Using Smartphone Sensing for Recognition of Game Player Attributes During Gameplay

Abstract

With the recent advances in smart sensing technologies, many people are now using smartphones to do routine tasks. In particular, the use of smartphones for gaming and infotainment purposes has increased significantly, as a result of which the gaming industry is expanding globally. Smartphones can now sense the interactions of a user with the device based on the embedded sensors. These human-smartphone interactions can be recognized using machine learning approaches to sense and automate/control different tasks being performed on the device. In particular, smartphone-based games have a lot of advantages, where the game controls can be seamlessly adjusted automatically based on the game player’s interaction to achieve a better gaming experience. In this regard, we propose a smartphone sensor-based method to recognize the attributes of a user, i.e., game player, during gameplay. The proposed scheme is based on the idea that different game players have their own ways of interacting with the device when playing games. The smartphone inertial sensors can be used to track these interactions and recognize different attributes (such as expertise level, gender, and identity) of a game player. This information can further be used in the games for adaptive control to maximize the user’s gaming experience. The proposed scheme is validated based on different experiments, and the overall average accuracy of 71.3% is achieved in the best case; thus, satisfactory results are achieved.